Microfinance institutions (MFIs) have introduced incentive pay schemes for their credit agents to induce information acquisition on borrowers. Bonuses linked to repayment are efficient for profit-oriented MFIs but insufficient for non-profit MFIs trying to reach very poor borrowers, when repayment and wealth are positively correlated. We show that no incentive scheme is consistent with this (non-verifiable) objective: Random audits on the share of very poor borrowers selected by the agent become necessary. Under the optimal contract, non-profit MFIs generally maximize the number of poor borrowers it services by cross-subsidization between very poor and less poor borrowers.

Though widespread, the practice of public subsidies for cultural activity lacks a rigorous and consistent economic rationale. We analyze a canonical market structure that characterizes much cultural activity: the competition of mass-produced goods with heterogeneous non-standardized goods that are imperfect substitutes. We analyze several types of market failure: uncertainty about preferences (we don’t know what we like, and we don’t know what we might like in the future); endogeneity of preferences (we like what our neighbors talk about, and we like what we’re accustomed to); and externalities associated with production (future production costs are determined by current production). The model provides a basis for cultural subsidies to promote social welfare and economic development.